pymorphy2-ua

Разбор украинского языка.ipynb

{"nbformat": 4, "cells": [{"outputs": [], "source": "from pymorphy2 import MorphAnalyzer, units\n\nmorph = MorphAnalyzer(\n    'dict-ua', \n    units = [\n        [\n            units.DictionaryAnalyzer(),\n#            units.AbbreviatedFirstNameAnalyzer(),\n#            units.AbbreviatedPatronymicAnalyzer(),\n        ],\n        units.NumberAnalyzer(),\n        units.PunctuationAnalyzer(),\n        [\n            units.RomanNumberAnalyzer(),\n            units.LatinAnalyzer()\n        ],\n#         units.HyphenSeparatedParticleAnalyzer(),\n        units.HyphenAdverbAnalyzer(),\n        units.HyphenatedWordsAnalyzer(),\n#         units.KnownPrefixAnalyzer(),\n        [\n            units.UnknownPrefixAnalyzer(),\n            units.KnownSuffixAnalyzer(),\n        ],\n        units.UnknAnalyzer(),\n    ]\n)", "execution_count": 1, "cell_type": "code", "metadata": {"trusted": true, "collapsed": true}}, {"outputs": [{"output_type": "execute_result", "metadata": {}, "execution_count": 2, "data": {"text/plain": "[Parse(word='\u0433\u0430\u0440\u043d\u0438\u0439', tag=OpencorporaTag('ADJF compb,nomn,sing,masc'), normal_form='\u0433\u0430\u0440\u043d\u0438\u0439', score=1.0, methods_stack=((DictionaryAnalyzer(), '\u0433\u0430\u0440\u043d\u0438\u0439', 73, 0),)),\n Parse(word='\u0433\u0430\u0440\u043d\u0438\u0439', tag=OpencorporaTag('ADJF compb,sing,masc,accs'), normal_form='\u0433\u0430\u0440\u043d\u0438\u0439', score=1.0, methods_stack=((DictionaryAnalyzer(), '\u0433\u0430\u0440\u043d\u0438\u0439', 73, 23),))]"}}], "source": "morph.parse('\u0433\u0430\u0440\u043d\u0438\u0439')", "execution_count": 2, "cell_type": "code", "metadata": {"trusted": true, "collapsed": false}}, {"outputs": [{"output_type": "execute_result", "metadata": {}, "execution_count": 3, "data": {"text/plain": "Parse(word='\u0433\u0430\u0440\u043d\u0456', tag=OpencorporaTag('ADJF compb,plur,nomn'), normal_form='\u0433\u0430\u0440\u043d\u0438\u0439', score=1.0, methods_stack=((DictionaryAnalyzer(), '\u0433\u0430\u0440\u043d\u0456', 73, 57),))"}}], "source": "morph.parse('\u0433\u0430\u0440\u043d\u0438\u0439')[0].inflect({'plur'})", "execution_count": 3, "cell_type": "code", "metadata": {"trusted": true, "collapsed": false}}, {"outputs": [], "source": "", "execution_count": null, "cell_type": "code", "metadata": {"trusted": true, "collapsed": true}}], "metadata": {"kernelspec": {"name": "python3", "display_name": "Python 3", "language": "python"}, "language_info": {"name": "python", "pygments_lexer": "ipython3", "version": "3.4.2", "mimetype": "text/x-python", "file_extension": ".py", "codemirror_mode": {"name": "ipython", "version": 3}, "nbconvert_exporter": "python"}}, "nbformat_minor": 0}